A digital approach in eliminating the higher order systematic effects

advertisement
A DIGITAL APPROACH IN ELIMINATING THE HIGHER
ORDER SYSTEMATIC EFFECTS IN PTOLEMY’S
GEOGRAPHIA LONGITUDE AND LATITUDE
DIFFERENCES
Angeliki Tsorlini and Evangelos Livieratos
Aristotle University of Thessaloniki
Faculty of Surveying Engineering
541 24 Thessaloniki, Greece
atsorlin@topo.auth.gr, livier@topo.auth.gr
Introduction
The interest in the geometric properties of historic maps has never been exhaustively and
continuously treated by analytical means, especially in the modern era of cartography. The
analytical treatment of the geometric background of early maps is an issue that today
attracts the attention it deserves, as a result of the challenging perspectives opened by new
digital technologies. These new technologies offer generously adequate processing tools
that allow diving into the world of the geometric origin and properties of historic cartographic representations and maps.
The core of this study concerns a two-dimensional spatial analysis of the field of
differences, testing various transformation functions in order to determine and eliminate
the systematic error pattern inherent in Ptolemy’s coordinates. The result, using ‘reduction’
methods in the comparison analysis (e.g. the concepts of the unit sphere, of the common
projective support) with all affined illustrations of the associated test, shows the pattern of
coordinate differences free of systematic effects up to the second order.
Ptolemaic reference system and coordinates
Ptolemy, in his Geographia, gives a list of geographic coordinates of spherical longitude
and latitude of almost ten thousand of point locations, on the earth surface, as known at
his times. These points are referred to geographic sites (i.e. towns, mountain picks,
river mouths, promontories and other) and their geographic coordinates are following the
known Ptolemaic reference system of parallels and meridians, the origin of which are
respectively close to actual Equator and close to the Canary Islands almost 25 degrees west
of the today’s origin at Greenwich (Figure 1).
Figure 1. The origin of parallels and meridians in Ptolemy’s Geographia
Coordinates in Ptolemy’s Geographia
In this paper, which is a part of a broader research carried out the last years by the
Cartography Group in the Faculty of Surveying Engineering at Thessaloniki, we focus our
interest on Ptolemy’s coordinates given in Geographia for the area of actual Greece, listed
mainly in Book III, Chapters XI to XV concerning Europe and Book V and Chapter II
concerning Asia. The tables referred to these chapters are Tables IX and X of Europe and
Table I of Asia, as it is shown in the Table below (Table 1), with the largest part of actual
Greece depicted in Table X (Figure 2). As it is obvious from the second figure, the world
of Ptolemy is classified in Regions, since each chapter is referred to one of them, giving by
this way the Atlas concept. The smaller the table is the more important and detailed the
region appears to be in Ptolemy’s Geographia.
Tables
Book III
Thrace
Chapter XI
Macedonia
Chapter XII
Epirus
Chapter XIII
Achaia
Chapter XIV
island Crete
Chapter XV
Ninth map of Europe
Tenth map of Europe
Book V
Asia properly so called (Asia
Minor)
Chapter II
First map of Asia
Table 1. Chapters and tables referred to Greece in Ptolemy’s Geographia
Figure 2. The ‘Tabulae’ in Ptolemy’s Geographia.
The digital process
According to the procedure we follow, we store, at the beginning, digitally the coordinates
for the area of interest from the different editions of Ptolemy’s Geographia we have, in a
database. In this case we are talking about almost 800 pairs of coordinates, 600 of them
refer to the actual territory of Greece. The editions, we use for this particular study are the
following five:
a. the Donnus Nicolaus Germanus mid-15th century manuscript of Ptolemy’s Geographia
as given in Codex Ebnerianus (Stevenson 1991: 92),
b. two 19th century editions by Nobbe (Leipzig 1843, 1966),
c. the Müller’s edition (Paris 1883),
d. the Vatopedion Codex (13th -14th century) and
e. the Marciana Codex (15th century).
Processing the Ptolemy’s Coordinates
The coordinates from the five sources, are independently and mutually checked and evaluated
in order to detect discrepancies in the point placement, and then projected onto a map with a
relevant graticule of parallels and meridians, all of them plotted in the same projection, e.g the
elementary geographic projection ( y  R, x  R ), assuming a unit radius reference sphere
( R  1 ) for the earth’s model. In this process maps are plotted from the coordinates and
Thrace
Macedonia
Asia
Minor
Epirus
Achaia
Crete
Figure 3. Coordinate plotting in geographic projection according to the snooped list of Ptolemy’s
coordinates and the actual borders of Greece. The point colours refer to the different regions, as
defined by Ptolemy.
by this way the locations of points are visualized, making easier the auto- and crosschecking of the values, the detection of the differences, the gross errors, the double values
and other displacements they may occur. In Figure 3 one of these maps is shown, in which
the toponyms are depicted concerning the regions that are referred totally or partially to the
actual territory of Greece. The yellow line shows in an approximate way, the actual borders
of modern Greece, while the different colours of the points are referred to the different
regions as described in Ptolemy’s Geographia.
Having concluded to an ‘accepted’ list of coordinates without gross errors, double values
or records or other apparently erroneous discrepancies in point placement, we start the
comparison of Ptolemy’s coordinates with their today’s counterparts. In order to perform
such a comparison and to identify the coincidence of places in Ptolemy’s era with their
today’s counterparts, we had to compare the toponyms of each area with the toponyms of
the corresponding area of actual territory of Greece confirming at the end the coincidence
of the with certainty known points in both cases, based mainly on relevant references in
historical and archeological sources. In the map of Figure 4, we can see on a modern map,
the places where some of Ptolemy’s toponyms are detected according to historical and
other sources.
Figure 4. Ptolemy’s toponyms on a modern map
Best fitting of Ptolemy’s representation to a modern map
The points, we have mentioned before, have great importance to the continuity of this work
because a set of them, properly distributed to the overall map space, is selected and
brought into one to one correspondence with the actual coordinates of the same set of
points in the modern map, after choosing a transformation system, in this case a 2nd order
polynomial transformation, involving a projection and an earth’s model. The result of the
best fitting of Ptolemy’s coordinates to the modern counterparts is shown in Figure 5. The
Ptolemy representation is georeferenced to actual geographic coordinates using almost 350
control points properly distributed in the area of Greece. Ptolemy’s graticule, extended
from 43˚ to 63˚ in longitude and from 33˚ to 45˚ in latitude, contrary to the geographic
graticule of the modern map, which extended from 18˚ to 30˚ in longitude and from 33˚ to
44˚ in latitude. In the resulting map (Figure 5), Ptolemy’s map of coordinates transforms
into the actual coordinates and the deformation appeared in Ptolemy’s graticule is obvious.
Figure 5. Second order polynomial best fitting of Ptolemy’s representation
in to a modern relevant map
The spatial distribution of differences in longitude and latitude
Previous research showed the order of magnitude of the longitude and latitude differences
of Ptolemy’s values from the today’s counterparts both in broader and local scale, but
without diving into a systematic geodetic approach on the issue. The problem is challenging and deserves a revisiting because of the advances of digital computational and
visualization technologies (informatics and infographics) which are massively available
today allowing new approaches and techniques in studying this extraordinary document of
our cartographic heritage as it is the Ptolemy’s Geographia.
Using the best fitting of Ptolemy’s representation to the modern map, we study also, the
spatial distribution of the differences in longitude and latitude induced after the comparison
of Ptolemy’s coordinates with their actual values. In the next two figures, which depict the
distribution of the differences in both cases (Figure 6 and 7), it is obvious that the
distribution is not the same in both cases.
As we can see below, the longitude differences vary from 25˚ at the west side of Greece to
31˚ at the east side whereas the latitude differences are of much smaller magnitude than
those of longitude and vary from -2˚ West to almost 1˚ East. These differences can be
easily explained by the fact that though latitudes are rather well defined, considering the
level of measuring accuracy at Ptolemy’s times, the longitudes suffered severe shortcomings which are due to the difficulties in measuring the time, which corresponds directly
to longitude. The longitude values given by Ptolemy are also strongly dependent upon the
distance from the Canaries eastwards.
Figure 6. The isolines of longitude differences, in degrees, between Ptolemy’s values
and their actual counterparts
Figure 7. The isolines of latitude differences, in degrees, between Ptolemy’s values
and their actual counterparts
Concluding remarks
The transformation of early maps into digital form and their comparison with modern maps
using new processing methods and technologies is of great importance for the study of the
geometric properties of early cartographic documents. Best fitting techniques are
appropriate in order to compare early cartographic representations with their modern
counterparts.
In this study particularly, the result of the two-dimensional spatial analysis of the field of
differences in Ptolemy’s coordinates shows the pattern of coordinate differences free of
systematic effects up to the 2nd order. This work is extended by testing also and some
higher order effects in order to get a better understanding of the whole process.
References
Boutoura C., E. Livieratos 2006. Some fundamentals for the study of the geometry of early
maps by comparative methods. e_Perimetron 1 (1): 60-70 [In digital form www.eperimetron.org].
Fischer J. 1991. Introduction. In E. L. Stevenson (ed.) Claudius Ptolemy, The Geography.
New York: Dover, 3-15.
Livieratos E. 2006. Graticule versus point positioning in Ptolemy cartographies.
e_Perimetron 1 (1): 51-59 [In digital form www.e-perimetron.org].
Livieratos E. 2006. On the Study of the Geometric Properties of Historical Cartographic
Representations. Cartographica (volume41, issue2) 165-175.
Marciana Codex 15th century (Codex CCCLXXXVIII) in facsimilae, Athens: Militos
Publ.
Müller C. 1883-1901. Claudii Ptolemaei Geographia. Bd. 1-2, Paris.
Nobbe C. (Hrsg., 1843-45). Klaudios Ptolemaios: Geographia. Leipzig. Reprint 1966.
Hildesheim.
Nordenskiöld A. E. 1978. Facsimile. Atlas to the early history of cartography. New York:
Dover.
Stevenson E. L., ed. 1991. Claudius Ptolemy, The Geography. New York: Dover.
Vatopedion Codex 13th century, in facsimilae, Athens: Militos Publ.
Download